Project · NHK Robocon · 2024-04
NHK Robocon 2024 Software Lead
Led the software team to build an autonomous mobile robot, developing a modular task architecture with Behavior Trees and ROS, and a 3D perception system.
Role & Responsibilities As the Software Lead for Kyoto University's NHK Robocon 2024 team, I spearheaded the development of an autonomous mobile robot capable of operating in uncertain environments. Key Technologies 3D Perception & Pose Estimation: Developed a system combining a YOLO model with a depth camera to compute the precise 3D global coordinates of objects. Modular Task Execution: Implemented a scalable architecture using Behavior Trees (instead of hard coded logic) combined with async communication in ROS for modular task development and debugging. YouTube Demo
#Robotics #ROS #Behavior Trees #Computer Vision #YOLO